DBScholar

Back to papers

FEC: Efficient Deep Recommendation Model Training with Flexible Embedding Communication

Summary: FEC uses embedding tiering and pre-fetching to cut embedding communication in EDRMs. AllReduce aggregates popular embeddings to mimic dense access; pre-fetching hides updates, delivering up to 6.65x embedding-communication and 2.42x throughput. (summarized by gpt-5-nano on Feb 09 2026)

Paper ID
6730
Venue
SIGMOD
Year
2023
Pagerank
5.3101785e-05
Overall Rank
9,164 | 37.13%
DOI
10.1145/3589310

Incoming Non-self Citations Over Time

Authors

BibTeX Citation

@inproceedings{ma_sigmod23,
        title = {{FEC: Efficient Deep Recommendation Model Training with Flexible Embedding Communication}},
        author = {Ma, Kaihao and Yan, Xiao and Cai, Zhenkun and Huang, Yuzhen and Wu, Yidi and Cheng, James},
        series = {{SIGMOD} '23},
        booktitle = {Proceedings of the {ACM} {SIGMOD} International Conference on Management of Data},
        publisher = {Association for Computing Machinery},
        doi = {10.1145/3589310},
        url = {https://dl.acm.org/doi/10.1145/3589310},
        year = {2023}
}

Incoming Citations (Sorted by Pagerank)

Showing 3 of 3 citing papers.

Previous Page 1 / 1 Next

Outgoing Citations (Sorted by Pagerank)

Showing 6 of 6 cited papers.

Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.

Previous Page 1 / 1 Next

Semantically Similar Papers